The effects of HIIT and MICT on body fat composition and cardiopulmonary fitness in adults: a meta-analysis of randomized controlled trials

Abstract

Objective

HIIT is a time-efficient aerobic exercise with potential for obesity prevention and control, as it achieves comparable or superior effects to traditional exercise in a shorter duration. This meta-analysis examines the differential effects of HIIT and MICT on body fat composition and cardiorespiratory fitness in adults, providing evidence for optimized exercise prescriptions.

Methods

Searched PubMed, Embase, Cochrane Library, and Web of Science for RCTs assessing HIIT cycling’s effects on fat reduction and cardiorespiratory fitness. Data were pooled using SMD in Review Manager 5.4 and Stata 18. Heterogeneity was evaluated via the I² statistic, with subgroup analyses stratified by population (disease, obese, sedentary).

Results

Nineteen RCTs were included. Compared to MICT, HIIT significantly reduced BMI in obese populations (SMD = -0.59 kg/m², 95% CI: -0.84 to -0.34, P < 0.001), but slightly increased BMI in disease populations (SMD = 0.40 kg/m², 95% CI: 0.03 to 0.77, P = 0.042). For cardiorespiratory fitness, HIIT outperformed MICT in improving VO₂peak in obese populations (SMD = 0.23 ml·kg⁻¹·min⁻¹, 95% CI: 0.01 to 0.45, P = 0.041) and sedentary populations (SMD = 6.21 ml·kg⁻¹·min⁻¹, 95% CI: 4.68 to 7.74, P < 0.001).

Conclusion

HIIT cycling demonstrates comparable efficacy to MICT for improving body composition, including body fat percentage, fat mass, and lean body mass in adults, while outperforming MICT in reducing BMI among obese individuals and enhancing cardiorespiratory fitness VO₂peak across obese and sedentary populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13102-025-01519-2.

Background

Contemporary socioeconomic transformation, coupled with the dual trends of increasingly refined dietary patterns and sedentary lifestyles, has exacerbated the global issue of energy metabolism imbalance among adults [1, 2]. According to the World Obesity Report 2024, the global population of obese adults is projected to rise from 524 million in 2010 to 1.13 billion by 2030, representing a 115% increase. Concurrently, non-communicable diseases associated with high BMI, including type 2 diabetes and cardiovascular diseases, account for 15% of deaths among individuals under 70 years of age globally, contributing to 17 million such fatalities annually [3, 4]. This positive energy balance-driven abnormal body composition, combined with declining cardiopulmonary function, forms a metabolic and cardiovascular disease vicious cycle, positioning obesity as a critical challenge with both medical and public health implications [5, 6].

Exercise intervention represents a cornerstone strategy for obesity prevention and management, with well-established efficacy supported by extensive clinical evidence. MICT traditionally considered the gold standard for fat-reduction exercise prescriptions, employs standardized protocols involving 30–60 min of aerobic exercise at 60–70% HRmax to effectively modify body fat distribution. However, modern lifestyle demands have increased the need for time-efficient exercise modalities. HIIT has emerged as a promising alternative, demonstrating unique potential for managing chronic conditions like diabetes and obesity through its characteristic protocol of 20-s bouts at 90% VO₂peak interspersed with 40-s recovery periods [7, 8]. Cycling presents an optimal exercise modality for obesity interventions, combining aerobic benefits with low joint impact. Compared to weight-bearing exercises like running, cycling reduces knee joint loading by approximately 40%, enhancing suitability for obese individuals and those with joint dysfunction [9, 10]. Recent evidence suggests that bicycle-based HIIT may offer superior advantages for fat reduction and cardiorespiratory improvement due to its time-efficient nature. However, systematic comparisons between HIIT cycling and traditional MICT remain insufficient in current literature. Notably, existing evidence exhibits inconsistencies regarding their efficacy, Sabag et al. reported that HIIT outperforms MICT in reducing liver fat [11]. While Khodadadi et al. meta-analysis found no significant difference in body fat percentage improvement between the two modalities [7]. For cardiopulmonary fitness, Chen et al. suggested HIIT is superior to MICT in enhancing VO₂peak, yet Redline et al. observed no such difference in clinical populations [12, 13]. Crucially, few studies have specifically focused on cycling, a low-impact modality leading to a lack of systematic comparisons between HIIT cycling and MICT cycling across diverse populations.

MICT remains the gold-standard exercise prescription for fat reduction. Its prolonged, steady-state exercise protocols at moderate intensity have demonstrated significant improvements in body composition and exercise capacity among obese populations. Conversely, HIIT has emerged as a time-efficient alternative, gaining widespread application in clinical prevention and rehabilitation settings. Current evidence suggests HIIT may surpass MICT in fat loss efficacy. To address existing research gaps, this study systematically reviewed literature from multiple databases to compare cycling-based HIIT and MICT interventions. We employed subgroup analyses to examine differential responses across clinical, obese, and sedentary populations, complemented by heterogeneity tests and sensitivity analyses to assess methodological biases. These findings will contribute evidence-based recommendations for exercise medicine while informing the development of population-specific training protocols through optimized intensity and duration parameters.

Methods

Search strategy

Through an analysis of existing literature on HIIT and MICT, relevant subject terms and their derivatives were identified to formulate a systematic search strategy. The search covered the period from January 1, 2000 to December 31, 2024. Following the predefined search formula, randomized controlled trials investigating HIIT and MICT were retrieved from databases including PubMed, Embase, Cochrane Library, and Web of Science. The detailed search syntax, including MeSH terms and free-text words was: (("High-Intensity Interval Training" OR "HIIT" OR "High intensity interval training") AND ("Moderate-Intensity Continuous Training" OR "MICT" OR "Moderate intensity continuous training") AND ("Cycling" OR "Bicycle") AND ("Fat Loss" OR "Body Composition" OR "BMI" OR "Body Fat Percentage") AND ("Cardiorespiratory Fitness" OR "VO₂peak") AND ("Randomized Controlled Trial")). Additional studies were identified through backward citation tracking of selected articles and related journals to ensure comprehensive coverage. This review adheres to the PRISMA guidelines. The study protocol has been registered in the PROSPERO (Registration number: CRD420251087399).

Inclusion and exclusion criteria

The inclusion and exclusion criteria were established according to PICOS principles [14]. Participants: Included studies involving adults aged 18–60 years, while excluding those outside this age range.Intervention: Included studies employing HIIT protocols, excluding those where intensity failed to reach 80–100% VO₂peak or maximum heart rate. Control: Included studies using MICT as control, excluding those with MICT durations < 20 min. Outcomes: Included studies reporting body composition measures (BMI, body fat, fat mass, lean mass) or VO₂peak, excluding those lacking these outcomes. Study design: Included only RCTs, excluding reviews, prospective cohort studies, incomplete datasets, conference abstracts, and proceedings. Two researchers independently conducted the screening process. Any discrepancies were resolved through discussion with a third researcher.

Data extraction

The literature retrieved from each database was imported into EndNote for duplicate removal. Two independent researchers conducted the screening process in two phases: (1) initial screening based on titles and abstracts, (2) full-text review of potentially eligible studies. For studies meeting the inclusion criteria, both researchers independently extracted study characteristics and primary outcome data. Any discrepancies in data extraction were resolved through consensus discussion with a third researcher.

Risk of bias assessment

Using the Cochrane Risk of Bias tool, and Review Manager 5.4. Detailed results of the literature quality evaluation are presented in Fig. 2, the evaluation focused on the following aspects: (1) Selection bias: whether the method of random sequence generation was used; (2) Allocation concealment: whether the allocation was effectively concealed; (3) Blinding: whether the subjects or investigators were blinded; (4) Data integrity: whether missing data was adequately reported and intention-to-treat analysis was applied; (5) Selective reporting: whether there was any selective reporting of outcomes; (6) Other biases: whether other factors contributed to potential bias. Primarily includes: Imbalances in baseline characteristics of participants; Potential biases in intervention implementation; Unreported confounding factors.

Fig. 2.

Risk of bias assessment

Statistical analysis

Data analysis was performed using Review Manager 5.4 and Stata 18.0 Effect sizes were calculated for pre- and post-intervention outcomes in both HIIT and MICT groups, including change values, using mean differences and standard deviations. For studies reporting only means and SE, SD were derived through conversion. SMD were used for outcomes with significant value differences or varying measurement units. Given potential variations in study populations and intervention protocols between trials, a random-effects model was employed for all meta-analyses. Heterogeneity was assessed using Cochrane's Q test and I² statistics. The following thresholds were applied: 0% ≤ I² < 50% indicated acceptable heterogeneity, while 50% ≤ I² < 100% suggested substantial heterogeneity. For any analyses showing heterogeneity (I² > 0%), sensitivity analyses were conducted to identify outlier studies and explore potential sources of heterogeneity through careful examination of the relevant publications.

Results

Literature selection

Our systematic search identified 1,326 articles related to HIIT, MICT, fat loss, and cardiopulmonary fitness from database searches, with an additional 841 records identified through citation searching. After removing duplicates, we screened 63 review articles in the first round. Application of inclusion/exclusion criteria through abstract and full-text review yielded 19 eligible studies for final inclusion. Figure 1 presents the complete study selection process.

Fig. 1.

Flow chart of literature retrieval

Literature information

This meta-analysis included 19 randomized controlled trials comparing HIIT and MICT interventions [15–32]. The final analysis comprised 527 participants who completed the trials (HIIT group: n = 267; MICT group: n = 260). All studies implemented cycling-based interventions lasting 2–24 weeks, with a frequency of 3–5 sessions per week. Table 1 summarizes the baseline characteristics of the included studies.

Table 1.

Basic information of the literature

Author	Age	M/F	Subjects	Type	Period	HIIT				MICT
						n	Intensity	Time	Frequency	n	Intensity	Time	Frequency
Atan, 2020 [15]	47.6 ± 8.7	0/40	fibromyalgia	cycle	6 weeks	19	4 min 80%−95%HRmax; 3 min 70%HRmax	25 min	5 times/week	19	65%−75% HRmax	45 min	5 times/week
Benham, 2021 [16]	18–40	0/30	PCOS	cycle	24 weeks	16	30 s 90%HRmax; 90 s Low intensity	20 min	3 times/week	14	50–60% HRmax	40 min	3 times/week
Boff, 2019 [17]	23.5 ± 6	8/10	diabetes	cycle	8 weeks	9	1 min 80%−85%HRmax; 4 min 50%HRmax	30 min	3 times/week	9	60%−65% HRmax	20–30 min	3 times/week
Cocks, 2017 [18]	25 ± 1	16/0	obesity	cycle	4 weeks	8	30 s 200% Wmax; 2 min 30W	10–15 min	5 times/week	8	60%VO2peak	40–60 min	5 times/week
Connolly, 2017 [19]	43.5 ± 7	0/36	obesity	cycle	12 weeks	15	30 s 30% HRmax; 20 s 50%−60% HEmax; 10 s 90% HRmax	15–25 min	3 times/week	15	Self-perception	30–50 min	3 times/week
Higgins, 2016 [19]	20.4 ± 1.5	0/52	obesity	cycle	6 weeks	23	30 s exhaustion exercise; 4 min restore	22.5–31.5 min	3 times/week	29	60%−70%HRmax	20–30 min	3 times/week
Hu, 2021 [20]	21.4 ± 1.4	0/33	obesity	cycle	12 weeks	15	4min90%VO2peak; 3 min rest	28–54 min	3 times/week	15	60% VO2peak	40–60 min	3 times/week
Keating, 2014 [21]	43 ± 8.3	5/21	obesity	cycle	12 weeks	11	VO2peak; 2–3 min Low intensity	14–18 min	3 times/week	11	50%−65% VO2peak	30–42 min	3 times/week
Mazurek, 2014 [22]	19.5 ± 0.6	0/46	obesity	cycle	8 weeks	24	10 s Sprint; 65%−75%HRmax	32 min	3 times/week	22	65%−75% HRmax	32 min	3 times/week
Middelbeek, 2021 [23]	48 ± 5	22/0	obesity	cycle	8 weeks	12	30 s exhaustion exercise; 4 min rest	27 min	3 times/week	10	60% VO2peak	40–60 min	3 times/week
Nie, 2017 [24]	21 ± 1.4	23/0	obesity	cycle	12 weeks	16	4 min 90%VO2peak; 3 min rest	300 KJ	3–4 times/week	14	60% VO2peak	300 KJ	3–4 times/week
Petrick, 2020 [25]	37.4 ± 15.1	23/0	obesity	cycle	6 weeks	12	30 s 170Wpeak; 2 min rest	10–15 min	3 times/week	11	60% Wpeak	30–40 min	5 times/week
Ram, 2020 [26]	28 ± 7	28/0	obesity	cycle	6 weeks	16	1 min 90%HRmax; 1 min 15%Wpeak	20 min	3 times/week	13	65%−75%HRmax	30 min	3 times/week
Saanijoki, 2015 [27]	48 ± 5	26/0	obesity	cycle	2 weeks	13	30 s 180% Maximum power; 4 min rest	18–27 min	3 times/week	13	60% VO2peak	40–60 min	3 times/week
Scott, 2019 [28]	29 ± 10.6	10/4	diabetes	cycle	7 weeks	7	1 min 100%VO2 peak; 1 min 50W	10–20 min	3 times/week	7	65% VO2peak	30–50 min	3 times/week
Sjoros, 2018 [29]	49 ± 4	16/10	diabetes	cycle	2 weeks	11	30 s exhaustion exercise; 4 min rest	18–27 min	3 times/week	10	60% VO2peak	40–60 min	3 times/week
Tsai, 2016 [30]	22.3 ± 5.9	40/0	Sedentary	cycle	6 weeks	20	3 min 80%VO2 peak; 3 min 40%VO2 peak	30 min	5 times/week	20	60% VO2peak	30 min	5 times/week
Way, 2020 [31]	55.9 ± 2.3	26/0	diabetes	cycle	12 weeks	12	4 min 90%VO2 peak; 5 min 50% VO2 peak	9 min	3 times/week	12	60% VO2peak	45 min	3 times/week
Winn, 2017 [32]	43.5 ± 11.5	18/0	obesity	cycle	4 weeks	8	4 min 80%VO2 peak; 3 min 50%VO2 peak	400 KJ	4 times/week	8	55% VO2peak	400 KJ	4 times/week

Risk of bias

As shown in Fig. 2, the risk of bias assessment revealed that seven studies reported specific random sequence generation methods, primarily utilizing computer-generated random numbers or random functions, while twelve studies merely mentioned randomization without methodological details. Regarding allocation concealment, only three studies implemented appropriate methods. Five studies employed single-blind designs, with none utilizing double-blind procedures. Two studies experienced participant attrition without reporting reasons or performing intention-to-treat analysis. While no selective reporting bias was detected, two studies demonstrated baseline imbalances and were thus rated as having other potential biases. The remaining studies provided insufficient information to assess other potential bias domains.

Results of meta-analysis

Subgroup analysis serves as a crucial method for examining heterogeneity in intervention effects when investigating the impact of high-intensity interval cycling on body composition and cardiopulmonary function across different populations. This approach provides detailed evidence to support precision exercise prescriptions. Our study performed stratified analyses of key indicators including BMI, body fat percentage, fat mass, lean mass, and VO₂peak to elucidate population-specific responses to this exercise modality. The subgroup analysis results are presented in Table 2.

Table 2.

Results of subgroup analysis

Subgroup	Number	SMD (95% CI)	P − value	Heterogeneity test
BMI (kg/m2)
Overall	17	− 0.24 (− 0.44, − 0.05)	0.012	I2 = 78%, P < 0.001
Disease population	5	0.38 (0.01, 0.75)	0.042	I2 = 73%, P = 0.005
Obese people	11	− 0.57 (− 0.81, − 0.32)	< 0.001	I2 = 75%, P < 0.001
Sedentary people	1	0.00 (− 0.62, 0.62)	> 0.999	−
Body fat (%)
Overall	11	− 0.16 (− 0.39, 0.06)	0.160	I2 = 22%, P = 0.24
Disease population	2	0.20 (− 0.32, 0.71)	0.450	I2 = 0%, P = 0.73
Obese people	9	− 0.25(− 0.51, 0.002)	0.052	I2 = 25%, P = 0.23
Fat mass (kg)
Overall	9	− 0.19 (− 0.43, 0.06)	0.134	I2 = 8%, P = 0.37
Disease population	1	0.25 (− 0.38, 0.89)	0.435	−
Obese people	8	− 0.26 (− 0.52, 0.003)	0.053	I2 = 0%, P = 0.48
Lean body mass (kg)
Overall	7	− 0.10 (− 0.38, 0.17)	0.461	I2 = 0%, P = 0.57
Disease population	2	− 0.13 (− 0.644, 0.39)	0.623	I2 = 20%, P = 0.26
Obese people	5	− 0.09 (− 0.42, 0.23)	0.575	I2 = 0%, P = 0.57
VO2peak (mL·kg⁻1·min⁻1)
Overall	19	0.25 (0.07, 0.43)	0.007	I2 = 82%, P < 0.001
Disease population	6	0.03 (− 0.31, 0.36)	0.878	I2 = 32%, P = 0.20
Obese people	12	0.23 (0.01, 0.45)	0.041	I2 = 68%, P < 0.001
Sedentary people	1	6.09 (4.55, 7.63)	< 0.001	−

GRADE assessment

To systematically verify the credibility of the intervention effect of HIIT compared with MICT on adult health outcomes, this study adopted the GRADE system for graded assessment. The assessment dimensions mainly included five dimensions: risk of bias, inconsistency, inappropriateness, imprecision, and publication bias. The specific results are shown in Table 3.

Table 3.

GRADE quality of evidence assessment

Outcome	No of participants (studies)	Certainty Assessment					Standardized Mean effect (95% CI)	GRADE*
		Risk of Bias	Inconsistency	Indirectness	Imprecision	Other
BMI	17 RCTs	Serious	Not serious	Not serious	Not serious	None	−0.25 kg/m2 [−0.45,−0.06]	⨁⨁⨁◯ MODERATE
Body Fat	11 RCTs	Serious	Serious	Not serious	Not serious	None	−0.17% [−0.39,0.06]	⨁⨁◯◯ LOW
Fat mass	9 RCTs	Serious	Serious	Not serious	Not serious	None	−0.18 kg [−0.43, 0.06]	⨁⨁◯◯ LOW
Lean body mass	7 RCTs	Serious	Not serious	Not serious	Not serious	None	−0.11 kg [−0.65, 0.38]	⨁⨁⨁◯ MODERATE
VO₂peak	19 RCTs	Serious	Not serious	Not serious	Not serious	None	0.26 mL·kg⁻1·min⁻1 [0.08, 0.44]	⨁⨁⨁◯ MODERATE

*Certainty of evidence according to Grading of Recommendations, Assessment, Development and Evaluations:

High: We are very confident in the estimated effect

Moderate: Our confidence in the estimated effect is moderate

Low: We have limited confidence in the estimated effect

Very low: We have very little confidence in the estimated efect

BMI

Figure 3 shows the meta-analysis results for BMI, comprising 17 pairwise comparisons. Significant heterogeneity was observed across studies (I² = 80.3%, P < 0.001), warranting a random-effects model. The analysis revealed that high-intensity interval cycling significantly reduced BMI compared to control conditions (SMD = −0.25 kg/m², 95% CI: −0.45 to −0.06, P = 0.014). Subgroup analyses demonstrated differential effects: BMI significantly increased in clinical populations (SMD = 0.40 kg/m², 95% CI: 0.03 to 0.77, P = 0.042), decreased in obese populations (SMD = −0.59 kg/m², 95% CI: −0.84 to −0.34, P < 0.001), and showed no significant change in sedentary populations (SMD = 0.00 kg/m², 95% CI: −0.62 to 0.62, P > 0.999).

Fig. 3.

Forest map of BMI subgroup analysis

Body fat

Figure 4 shows the results of the meta-analysis of body fat. The heterogeneity among the literatures is relatively small (I² = 29.6%, P = 0.163), and a fixed-effects model was used for the analysis. The results of the meta-analysis showed that compared with the control group, high-intensity intermittent cycling had no significant effect on body fat percentage (SMD = −0.17%, 95% CI: −0.39 to 0.06, P = 0.160). The results of subgroup analysis showed that high-intensity intermittent cycling had effects on disease (SMD = 0.20%, 95% CI: −0.31 to 0.72, P = 0.450) and obesity (SMD = −0.26%, 95% CI: −0.51 to 0.00, P = 0.052) There was no significant effect on the body fat percentage of the population.

Fig. 4.

Forest map of body fat subgroup analysis

Fat mass

Figure 5 shows the meta-analysis results for fat mass, incorporating 9 pairwise comparisons. The analysis revealed low heterogeneity across studies (I² = 7.9%, P = 0.369), justifying the use of a fixed-effects model. No significant difference was found between high-intensity interval cycling and control conditions (SMD = −0.18 kg, 95% CI: −0.43 to 0.06, P = 0.134). Subgroup analyses indicated non-significant effects for both clinical populations (SMD = 0.25 kg, 95% CI: −0.38 to 0.89, P = 0.435) and obese populations (SMD = −0.26 kg, 95% CI: −0.52 to 0.00, P = 0.575).

Fig. 5.

Forest map of fat mass subgroup analysis

Lean body mass

Figure 6 shows the meta-analysis results for lean body mass, comprising data from seven studies. The analysis demonstrated minimal heterogeneity (I² = 0.0%, P = 0.511), supporting the use of a fixed-effect model. The pooled results indicated no significant effect of high-intensity interval cycling on lean body mass compared to control conditions (SMD = −0.11 kg, 95% CI: −0.38 to 0.16, P = 0.461). Subgroup analyses revealed non-significant effects for both clinical populations (SMD = −0.14 kg, 95% CI: −0.65 to 0.38, P = 0.623) and obese populations (SMD = −0.10 kg, 95% CI: −0.42 to 0.22, P = 0.541).

Fig. 6.

Forest map of Lean body mass subgroup analysis

VO₂peak

Figure 7 presents the meta-analysis results for VO₂peak, incorporating 19 pairwise comparisons. Substantial heterogeneity was observed across studies (I² = 83.0%, P < 0.001), necessitating a random-effects model. The pooled analysis demonstrated that HIIT significantly improved VO₂peak compared to control conditions (SMD = 0.26 mL·kg⁻¹·min⁻¹, 95% CI: 0.08 to 0.44, P = 0.007). Subgroup analyses revealed significant improvements in obese populations (SMD = 0.23 mL·kg⁻¹·min⁻¹, 95% CI: 0.01 to 0.45, P = 0.041) and sedentary populations (SMD = 6.21 mL·kg⁻¹·min⁻¹, 95% CI: 4.68 to 7.74, P < 0.001), while showing no significant effect in clinical populations.

Fig. 7.

Forest map of VO₂peak subgroup analysis

Sensitivity analysis

To assess the robustness of the meta-analysis findings, we performed a sensitivity analysis using the leave-one-out method for outcome measures with high heterogeneity. The results indicated that sequentially excluding any single study did not substantially alter the direction or statistical significance of the pooled effect size, indicating that the findings are robust. Furthermore, although the exclusion of individual studies occasionally reduced the I² value, the overall impact on the heterogeneity estimate was minimal, confirming that no single study exerted a dominant influence on the analysis. The specific results of the sensitivity analysis can be found in the supplementary materials.

Heterogeneity analysis

Heterogeneity in this study was assessed using Cochran's Q test, which indicated significant heterogeneity for both BMI (I² > 50%) and VO₂peak (I² > 50%). Subgroup analysis of BMI revealed statistically significant differences in effect sizes across population subgroups (p = 0.009), confirming that participant characteristics are a major source of heterogeneity. Notably, heterogeneity was highest in the obese subgroup (I² = 75.2%, p < 0.001), likely attributable to variations in obesity classification and baseline health status among the included samples. In the VO₂peak subgroup analysis, heterogeneity was moderately low in the disease population (I² = 31.7%, p = 0.198), suggesting relatively consistent effect directions and magnitudes across studies. In contrast, high heterogeneity persisted in the obese subgroup (I² = 68.2%, p < 0.001), again possibly due to differences in obesity severity and baseline health. Funnel plots supporting this heterogeneity assessment are provided in the supplementary materials.

Discussion

This systematic review evaluated the effects of cycling-based HIIT on body composition and cardiopulmonary fitness in adults. Meta-analysis demonstrated that HIIT cycling significantly reduced BMI and increased VO₂peak compared to control conditions. However, no significant effects were observed for body fat percentage, fat mass, or lean mass. Subgroup analyses revealed population-specific responses, indicating that the therapeutic efficacy of HIIT cycling varies across different demographic groups. The results revealed that HIIT substantially decreased BMI in obese individuals but increased BMI in those with preexisting medical conditions. This divergence may result from metabolic compensatory mechanisms in pathological states [33]. Notably, high heterogeneity was observed in some analyses, which may stem from three main factors. Firstly, variations in baseline characteristics of participants, BMI ranges of obese individuals (28–35 kg/m²) and disease durations of clinical populations (1–10 years) varied significantly across studies, potentially affecting intervention responses. Secondly, differences in HIIT protocol design intensity (80%–100% VO₂peak), interval duration (10 s–4 min), and recovery methods (low-intensity cycling/rest) differed. Finally, confounding dietary factors, some studies failed to strictly control diet, which may interact with exercise effects and further increase heterogeneity.

Specifically, in diabetic patients, chronic hyperglycemia may exacerbate inflammatory stress induced by HIIT, leading to increased muscle protein breakdown; Meanwhile, some patients exhibit increased compensatory eating after exercise, ultimately resulting in a slight increase in BMI [34]. In patients with fibromyalgia, the disease itself is accompanied by chronic inflammation and muscle dysfunction, and the high-intensity nature of HIIT may aggravate muscle damage and cause tissue edema [35]; additionally, patients with low exercise tolerance fail to achieve the expected energy expenditure, leading to no decrease or even a slight increase in BMI [36]. In obese individuals, HIIT likely reduces BMI by enhancing fat oxidation via AMPK pathway activation while concurrently increasing resting metabolic rate [37, 38]. Notably, HIIT cycling had no significant effect on BMI in sedentary individuals, suggesting that their baseline body composition exhibits lower responsiveness to HIIT and may require extended intervention durations or complementary exercise modalities.

Compared to MICT, HIIT cycling did not significantly alter body fat percentage or total fat mass in adults. Subgroup analyses similarly revealed no substantial changes in either clinical or obese populations. This null effect may be attributed to HIIT unique energy expenditure profile [39, 40]. While HIIT enhances excess EPOC, the ≤ 12-week intervention duration in most included studies may have been insufficient to induce meaningful adipose tissue redistribution. Additionally, uncontrolled dietary variables in some studies may have obscured potential HIIT induced modifications in body composition. Both primary and subgroup analyses indicated stable lean body mass indices, implying that HIIT hypertrophic effects are less pronounced than MICT. The marginal reduction in lean mass observed in clinical populations may reflect disease-related muscle catabolism potentiated by exercise-induced metabolic stress [41, 42]. However, the limited sample sizes and available literature necessitate further investigation to elucidate differential exercise modality effects on lean tissue preservation. This study demonstrates that HIIT cycling significantly improves VO₂peak through multiple physiological mechanisms. First, HIIT induces cardiovascular adaptive remodeling: in obese individuals, it enhances cardiac output by improving HRV and vascular endothelial function [43]. Sedentary individuals, who typically exhibit reduced baseline cardiopulmonary reserves, experience acute hypoxic stress during HIIT that stimulates erythropoietin release and augments oxygen transport capacity [44]. Second, HIIT promotes mitochondrial biogenesis via intermittent hypoxia-reoxygenation cycles, which activate the PGC-1α/NRF-1 pathway, increase skeletal muscle mitochondrial density, and improve oxidative phosphorylation efficiency [45]. Notably, VO₂peak remained unchanged in clinical populations, potentially due to disease-imposed limitations on cardiopulmonary reserve or sample heterogeneity [46].

This study examining HIIT versus MICT reveals both confirmatory and novel findings relative to existing literature. Our comparative analysis focuses on two key dimensions: body composition regulation and cardiopulmonary fitness enhancement. Regarding BMI regulation, HIIT significantly reduced BMI in obese participants (SMD = −0.57, P < 0.001), aligning with Programa et al. [47] findings (SMD = −0.49) but demonstrating greater effect size. This enhanced efficacy may stem from improved exercise adherence facilitated by cycling's low-impact nature [47]. Conversely, we observed increased BMI in clinical populations (SMD = 0.38, P = 0.042), consistent with Jing-Xin Liu et al. [48] observations in type 2 diabetes patients. This phenomenon may reflect transient fluid retention from exercise-induced cortisol secretion, suggesting that chronic disease states may attenuate HIIT body composition benefits [48].For body fat metrics, our results showing no significant HIIT and MICT difference corroborate Tucker et al. [49] conclusion that ≤ 12-week interventions may be insufficient to induce adipocyte apoptosis or sustained lipase upregulation [49]. However, these findings contrast with Yuan et al. [50] potentially due to their stricter dietary controls which may better reveal HIIT lipolytic potential [50]. Concerning cardiopulmonary fitness, HIIT significantly improved VO₂peak in obese and sedentary cohorts, supporting Russomando et al. [51] framework of superior HIIT adaptation through enhanced ventricular ejection fraction and skeletal muscle capillarization [51]. Notably, clinical populations showed no VO₂peak improvement, consistent with Niyazi et al. [52] findings that disease impaired cardiopulmonary reserve may limit HIIT efficacy, underscoring the need for careful intensity prescription in these populations [52].

Clinical significance

This study offers clinically relevant guidance for prescribing targeted exercise programs to key subgroups. For individuals with obesity, HIIT protocol on a cycle ergometer—consisting of 4 min intervals at 90%VO₂peak followed by 3 min active recovery at 50%VO₂peak, performed 3–4 times/week for over 12 weeks is recommended. This regimen can effectively lower BMI and increase VO₂peak, especially when combined with dietary control to enhance body fat reduction. For individuals with obesity and comorbidities such as diabetes or fibromyalgia, direct implementation of a standard HIIT protocol is not advised. Pre-exercise risk screening should be performed first. A modified, low intensity HIIT program performed three times weekly, with an initial shortened intervention period of six weeks, is preferable. Close monitoring of changes in BMI and disease specific markers is essential to avoid exercis related adverse events. For sedentary individuals, a progressive approach is recommended, beginning with lower intensity HIIT (e.g., 3 min at 75%VO₂peak/3 min at 50%VO₂peak) five times/week for over 12 weeks. Extending the intervention to 24 weeks can further amplify improvements in BMI and VO₂peak. Adherence may be supported by incorporating low impact activities such as walking.

Limitations

Firstly, despite the large sample size included in this study, gender differences in exercise responses were not accounted for, and stratified analyses for male and female participants were not performed separately. Secondly, the composition of included participants was heterogeneous, with no consideration of the impact of age stratification or comorbidities on the overall meta-analysis results. Finally, notable variations existed between the HIIT and MICT intervention protocols included in the study. Although random-effects models were employed alongside sensitivity analyses and publication bias assessments, such discrepancies remain non-negligible. Future research should investigate participant characteristics, such as age, sex, and baseline health status in greater detail, with clinical trials designed to explicitly report subgroup differences. For HIIT, comparative studies are needed to examine how variations in protocol design, including intervention duration, work-to-rest ratios, and exercise modality, affect outcomes, thereby helping to standardize core intervention parameters. Further exploration is also required to identify optimal strategies for maximizing fat loss and improving cardiorespiratory fitness in adults. Integrating molecular and physiological approaches to uncover the mechanisms underlying population-specific responses will provide a stronger theoretical foundation for personalizing and optimizing future interventions.

Conclusion

HIIT cycling demonstrates comparable efficacy to MICT for improving body composition, including body fat percentage, fat mass, and lean body mass in adults, while outperforming MICT in reducing BMI among obese individuals and enhancing cardiorespiratory fitness VO₂peak across obese and sedentary populations. Considering the advantages and time efficiency of HIIT, it offers exercise options for adults in promoting fat reduction and increasing cardiopulmonary endurance.

Abbreviations

HIIT

High-intensity interval training

MICT

Moderate-intensity continuous training

BMI

Body mass index

SMD

Standardized mean differences

RCTs

Randomized controlled trials

VO₂peak

Peak oxygen uptake

PCOS

Polycystic ovary syndrome

Authors’ contributions

The authors confrim contribution to the paper as follows: Conceptualization: H.C., H.L.; Data curation: S.S., M.T., B.L.; Formal Analysis:H.C, B.L.; Writing-original draft、Writing- review & editing: S.S.. H.S.. All authors reviewed the results and approved the final version of the manuscript.

Funding

The Youth Project of the 2024 Beijing Digital Education Research Project (Number: BDEC2024QN048); The 2024 Chaoyang District Education Science “14th Five-Year Plan” Project of Beijing (Number:2023ZX116).

Data availability

The datasets generated and/or analysed during the current study are not publicly available at this stage, as they contain private information of participants involved in the original included trials to protect their confidentiality. The original data will be made publicly accessible after the manuscript is formally published, in compliance with ethical and privacy protection guidelines. In the interim, reasonable requests for access to the datasets can be directed to the corresponding author, and access will be granted upon verification of legitimate research purposes and adherence to participant privacy protection protocols. All analytical methods and key parameters used in this meta-analysis are fully described in the “ Methods” section of the manuscript to ensure reproducibility.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.